A flexible web position positioning calibration method and warehouse management system

Abstract

The present application relates to the field of warehouse management, and particularly relates to a flexible coiled material position positioning calibration method and a warehouse management system; comprising the following sequentially executed steps: S1: a physical information neural network with added physical constraints suitable for the environment of the coiled material to be stored is used to predict the radius of the coiled material, and a predicted radius is obtained; S2: a positioning algorithm is used to obtain the initial position coordinates of the coiled material to be stored, and the coiled material to be stored is mapped to a digital twin platform in combination with the predicted radius; S3: the initial position of the coiled material to be stored is judged for legality according to a preset physical rule, and if the legality judgment is not passed, step S4 is executed; S4: a slime algorithm with initial position priori is used to search for a target position conforming to the physical rule near the initial position, and S5: the coiled material to be stored is remapped to the digital twin platform according to the coordinates of the target position.

Description

Technical Field

[0001] This invention relates to the field of warehouse management, specifically to a method for calibrating the position of flexible roll materials and a warehouse management system. Background Technology

[0002] Flexible roll material storage typically involves high-density storage and turnover within rack-mounted storage locations. The processes of receiving, transferring, inventory counting, and picking heavily rely on the precise spatial location of the roll material. In recent years, warehouse management solutions combining RFID automatic identification and digital twin 3D visualization have been gradually adopted. These solutions use readers to collect tag signals for batch identification and then present the spatial relationship between the roll material and the racks in 3D on a digital twin platform, providing support for location management, picking paths, and operational scheduling.

[0003] However, in real-world shelving environments, positioning results based on signal characteristics such as RSSI are easily affected by multipath reflections, metal shelving structures, label posture and occlusion, differences in reader deployment, and dynamic interference on-site, leading to initial coordinate offsets and fluctuations. If these initial coordinates are directly mapped to a digital twin platform, inconsistencies with the physical realities of the scene often occur. For example, the roll material model may appear to "float" relative to its support surface, or it may "collide" or "penetrate" adjacent roll materials or shelving components. This distorts the digital twin display and subsequent business decisions, and may even lead to risks such as incorrect storage locations and incorrect picking.

[0004] On the other hand, the occupancy boundary and collision determination of flexible roll materials in three-dimensional space are highly dependent on geometric parameters such as roll radius (roll diameter). Existing systems mostly rely on manual measurement or empirical estimation of roll diameter, or simply perform regression fitting based on inventory information. This makes it difficult to take into account the generalization ability and physical consistency under different roll material types and length / weight differences, which can easily lead to the accumulation of geometric modeling errors and further amplify the inconsistency between "positioning coordinates-geometric shape-collision relationship".

[0005] To address the aforementioned issues, existing correction methods often employ threshold rule correction or general intelligent optimization algorithms to perform secondary searches of coordinates. However, without a feasibility determination mechanism tightly coupled with the physical rules of the warehousing scenario, and lacking a calibration strategy that involves "small-scale, minimal modifications" around the initial predicted coordinates, the calibration points are prone to deviating from the actual storage location, converging to positions with no business significance, or, although the error is reduced, still failing to pass the detection of constraints such as suspension / collision. This makes it difficult to meet the engineering requirements of "reliable, interpretable, and implementable" roll material warehousing positioning. Summary of the Invention

[0006] The purpose of this invention is to provide a positional calibration method for flexible roll materials and a warehouse management system to improve calibration accuracy.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A method for calibrating the position of a flexible roll material includes the following steps performed sequentially:

[0009] S1: Obtain the physical information neural network, obtain training samples, pre-train the physical information neural network, add physical constraints adapted to the environment of the warehouse roll material to the loss function, obtain the total loss function, use the total loss function to backpropagate and adjust the trained physical information neural network, obtain the prediction network, use the prediction network to predict the roll radius of the warehouse roll material, and obtain the prediction radius.

[0010] S2: Using a positioning algorithm, the initial position coordinates of the incoming roll material are obtained and combined with the predicted radius to map the incoming roll material to the digital twin platform;

[0011] S3: Perform a validity check on the initial position of the incoming roll material according to the preset physical rules. If the validity check fails, proceed to step S4; otherwise, proceed to step S5.

[0012] S4: The slime mold algorithm with initial position prior is used to search for target positions that conform to physical rules near the initial position. The specific steps are as follows:

[0013] Each slime mold location represents a suitable location for placing the incoming roll material. Based on the traditional slime mold algorithm, the local search is modified by incorporating an initial location prior, as shown in the following formula:

[0014] ;

[0015] ;

[0016] ;

[0017] in, The maximum number of iterations, Indicates the first iteration Indicates taking 0 to... Random numbers between For the first The location of the slime mold, The initial two-dimensional coordinates representing the location of the incoming roll material. This represents the mixed coefficient vector of the slime mold algorithm in the local search mode. This represents the convergence factor that varies linearly with iteration. , , , This represents the initial three-dimensional coordinates of the incoming roll material obtained by the positioning algorithm;

[0018] The fitness value of the population is calculated iteratively using the fitness function of this slime mold algorithm.

[0019] Repeat the calculation of the fitness function and update the population until the preset number of iterations is reached, and output the optimal result of the slime mold algorithm, which is the target location of the rolled material to be put into storage.

[0020] S5: Remap the incoming roll material to the digital twin platform based on the coordinates of the target location.

[0021] Preferably, in step S1, the training sample includes the type, length and weight of the stock rolls, and the physical constraints include weight ratio consistency, monotonicity prior and second-order curvature smoothness prior.

[0022] The calculation steps for the total loss function in step S1 are as follows:

[0023] S1-1: The loss function for the consistency of this weight ratio is:

[0024] ;

[0025] in, Indicates the quality of the roll material. Indicates the number of samples. Indicates the predicted radius. Indicates the radius of the roll material in the warehouse;

[0026] S1-2: The monotonicity prior is based on the length of the incoming roll material. After sorting, the corresponding prediction radius It should be monotonous and undiminished; One sample, for the length of the roll material entering the warehouse. After sorting in ascending order, we obtain the sorted index. :

[0027] ;

[0028] The loss function for monotonic priors is:

[0029] ;

[0030] in, The radius corresponding to the length of the incoming roll material after sorting.

[0031] S1-3: The loss function for the second-order curvature smoothing prior is:

[0032] ;

[0033] S1-4: The total loss function is expressed by the following formula:

[0034] ;

[0035] ;

[0036] in, For the first The true value of the radius of each training sample. To monitor errors, , , These are the weighting coefficients for each loss function.

[0037] Preferably, in step S3, the specific steps for determining the legality of the initial location of the incoming roll material are as follows:

[0038] S3-1: Perform suspension detection on the incoming roll material based on Euclidean spatial measurement. The bounding box of the incoming roll material in the digital twin platform is:

[0039] ;

[0040] The center of the bottom support point of the incoming roll material is The coordinates of the two ends of the shelf in the warehouse where the incoming roll material is located are respectively , ,

[0041] in, This represents the horizontal coordinate of the shelving unit in this storage location. and This indicates that the incoming roll material is in the digital twin platform. Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis;

[0042] S3-2: Project the bottom center point of the incoming roll material onto the shelf. Plane, projection point is ;

[0043] Define the shelf vector and the projection point vector, and calculate the projection ratio of the projection point vector onto the shelf vector using the following formula:

[0044] ;

[0045] ;

[0046] like The projection point will fall inside the shelf. Indicates the threshold height for shelf support. This is an empirical value; otherwise, use the following formula to calculate the ratio of the vertical height of the bottom center of the incoming roll material to the shelf height:

[0047] ;

[0048] like If the incoming roll material is supported by the shelf, then proceed to step S3-3;

[0049] S3-3: The following formula is used to determine whether the incoming roll material is supported by other roll materials:

[0050] ;

[0051] ;

[0052] ;

[0053] in, This indicates the first [unit / location] of the storage location. The size of the wrapping box for each roll of material. This indicates the first [unit / location] of the storage location. The center point of the roll material is at Projected coordinates of the plane The radius of the roll of material to be stored;

[0054] Calculate the horizontal distance between the bottom center of the incoming roll material and other roll materials:

[0055] ;

[0056] in, and Is the incoming roll material in The coordinates of the projection point on the plane, This indicates that the incoming roll material is related to the first item in the warehouse location. Each roll of material Distance between points projected onto the plane;

[0057] If satisfied and , This is the tolerance value for horizontal difference. The threshold height for the support structure of the incoming roll material. If the value is empirical, then it is considered that the incoming roll material is stored in the warehouse location. Supported by the roll material, without any suspension, proceed to step S3-4 to perform multi-level collision detection;

[0058] S3-4: Using an axis-aligned bounding box algorithm for rapid detection, the index of the roll that collides with the incoming roll is selected from all rolls in this storage location:

[0059] ;

[0060] ;

[0061] in, and These respectively represent the incoming roll material and the first one in the storage location. A box surrounding a roll of material. For roll materials that need to be further checked for collisions after rapid initial screening by the axis alignment box;

[0062] like If the length is not empty, the following formula is used to determine... Are there any rolls in the collection that collided with the incoming roll?

[0063] ;

[0064] ;

[0065] in, Indicates the center of the roll material is at The projection on the plane, then and These respectively indicate the condition of the incoming roll material and the roll material after initial screening. Projection on a plane Indicates the radius of the roll material, then and These represent the radii of the incoming roll material and the roll material after initial screening, respectively.

[0066] For each item It is necessary to determine whether the following collision determination conditions are met:

[0067] ;

[0068] in, This is the tolerance value for the radius.

[0069] Preferably, in step S4, the fitness function of the slime mold algorithm is:

[0070] ;

[0071] ;

[0072] ;

[0073] ;

[0074] in, The fitness value represents the slime mold. Indicates the first The location of each slime mold, Indicates the first The distance between the location of each slime mold and the predicted location of the incoming roll material is used to encourage slime molds to move closer to the coordinates of the incoming roll material. This indicates whether the location of the slime mold is surrounded by other rolls of material in that storage location. If so, the index of the roll of material in that location is recorded in the set. middle, Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the radius of the roll material entering the warehouse. Indicates the magnitude of suspension error. Indicates the first The slime mold represents the roll material entering the warehouse. The height difference between the lowest point of the coordinate system and the shelf. This indicates the shelving in this warehouse location. coordinate, Indicates the first The slime mold represents the roll material entering the warehouse. The lowest point of the coordinates and the first storage location One roll of material The height difference of the highest points of the coordinates Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the location of the storage unit. The radius of each roll of material, Indicates the magnitude of the collision error. Indicates the first The boundary of the roll material represented by each slime mold in the warehouse is related to the first location in this warehouse. The edge bonding condition of each roll material.

[0075] A warehouse management system for flexible roll materials includes a roll material radius prediction module, a roll material positioning module, a roll material positioning validity judgment module, a roll material positioning correction module, and a roll material position display module;

[0076] The roll radius prediction module is used to acquire physical information neural network and training samples, and to pre-train the physical information neural network using the training samples. Physical constraints adapted to the environment of the roll material entering the warehouse are added to the loss function of the physical information neural network to obtain the total loss function. The total loss function is used to backpropagate and adjust the trained physical information neural network to obtain the prediction network. The prediction network is used to predict the roll radius of the roll material entering the warehouse to obtain the predicted radius.

[0077] The roll material positioning module is used to obtain the initial position coordinates of the inbound roll material using a positioning algorithm and combine the predicted radius to map the inbound roll material onto the roll material position display module;

[0078] The roll material positioning legality judgment module is used to judge the legality of the initial position of the incoming roll material according to preset physical rules;

[0079] The roll material positioning correction module uses a slime mold algorithm with an initial position prior to search for a target position that conforms to physical rules near the initial position. The specific operation process is as follows:

[0080] Each slime mold location represents a suitable location for placing the incoming roll material. Based on the traditional slime mold algorithm, the local search is modified by incorporating an initial location prior, as shown in the following formula:

[0081] ;

[0082] ;

[0083] ;

[0084] in, The maximum number of iterations, Indicates the first iteration Indicates taking 0 to... Random numbers between For the first The location of the slime mold, The initial two-dimensional coordinates representing the location of the incoming roll material. This represents the mixed coefficient vector of the slime mold algorithm in the local search mode. This represents the convergence factor that changes linearly with iteration.

[0085] The initial three-dimensional coordinates of the incoming roll material are converted into two-dimensional coordinates from the roll material location display module.

[0086] The fitness value of the population is calculated iteratively using the fitness function of this slime mold algorithm.

[0087] Repeat the calculation of the fitness function and update the population until the preset number of iterations is reached, and output the optimal result of the slime mold algorithm, which is the target location of the rolled material to be put into storage.

[0088] The roll material location display module is used to display the target location of the incoming roll material on the system front end.

[0089] Preferably, the training sample includes the type, length and weight of the stock rolls, and the physical constraints include weight ratio consistency, monotonicity prior and second-order curvature smoothness prior;

[0090] The steps for calculating the total loss function are as follows:

[0091] The loss function for the consistency of the weight ratio is:

[0092] ;

[0093] in, Indicates the quality of the roll material. Indicates the number of samples. Indicates the predicted radius. Indicates the radius of the roll material in the warehouse;

[0094] The monotonicity prior is the length of the incoming roll material. After sorting, the corresponding prediction radius It should be monotonous and undiminished; One sample, for the length of the roll material entering the warehouse. After sorting in ascending order, we obtain the sorted index. :

[0095] ;

[0096] The loss function for monotonic priors is:

[0097] ;

[0098] in, The radius corresponding to the length of the incoming roll material after sorting.

[0099] The loss function for the second-order curvature smoothing prior is:

[0100] ;

[0101] in, The radius corresponding to the length of the incoming roll material after sorting.

[0102] The total loss function is expressed by the following formula:

[0103] ;

[0104] ;

[0105] in, For the first The true value of the radius of each training sample. To monitor errors, , , These are the weighting coefficients for each loss function.

[0106] Preferably, the specific steps of the roll material positioning legality judgment module to judge the legality of the initial position of the incoming roll material are as follows:

[0107] The floating detection of the incoming roll material was performed based on Euclidean spatial measurement. The bounding box of the incoming roll material in the digital twin platform is as follows:

[0108] ;

[0109] The center of the bottom support point of the incoming roll material is The coordinates of the two ends of the shelf in the warehouse where the incoming roll material is located are respectively , ,

[0110] in, This represents the horizontal coordinate of the shelving unit in this storage location. and This indicates that the incoming roll material is in the digital twin platform. Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis;

[0111] Project the center point of the bottom of the incoming roll of material onto the shelf. Plane, projection point is ;

[0112] Define the shelf vector and the projection point vector, and calculate the projection ratio of the projection point vector onto the shelf vector using the following formula:

[0113] ;

[0114] ;

[0115] like The projection point will fall inside the shelf. This indicates that the shelf support height threshold is an empirical value; otherwise, the ratio of the vertical height of the bottom center of the incoming roll material to the shelf height is calculated using the following formula:

[0116] ;

[0117] like If the incoming roll material is supported by the shelf, then proceed to step S3-3;

[0118] The following formula is used to determine whether the incoming roll material is supported by other roll materials:

[0119] ;

[0120] ;

[0121] ;

[0122] in, This indicates the location of the storage unit. The size of the wrapping box for each roll of material. This indicates the location of the storage unit. The center point of the roll material is at Projected coordinates of the plane The radius of the roll of material to be stored;

[0123] Calculate the horizontal distance between the bottom center of the incoming roll material and other roll materials:

[0124] ;

[0125] in, and Is the incoming roll material in The coordinates of the projection point on the plane, This indicates that the incoming roll material is related to the first item in the warehouse location. Each roll of material Distance between points projected onto the plane;

[0126] If satisfied and , This is the tolerance value for horizontal difference. If the threshold for the support height of the incoming roll material is taken as an empirical value, then it is considered that the incoming roll material is supported by the first unit of the storage location. Supported by the roll of material, it did not become suspended.

[0127] A bounding box algorithm with axis alignment is used for rapid detection to select the index of the roll that may collide with the incoming roll from all rolls in this storage location:

[0128] ;

[0129] ;

[0130] in, and These respectively represent the incoming roll material and the first one in the storage location. A box surrounding a roll of material. For roll materials that need to be further checked for collisions after rapid initial screening by the axis alignment box;

[0131] like If the length is not empty, the following formula is used to determine... Are there any rolls in the collection that collided with the incoming roll?

[0132] ;

[0133] ;

[0134] in, Indicates the center of the roll material is at The projection on the plane, then and These respectively indicate the condition of the incoming roll material and the roll material after initial screening. Projection on a plane Indicates the radius of the roll material, then and These represent the radii of the incoming roll material and the roll material after initial screening, respectively.

[0135] For each item It is necessary to determine whether the following collision determination conditions are met:

[0136] ;

[0137] in, This is the tolerance value for the radius.

[0138] Preferably, the fitness function of this slime mold algorithm is:

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] in, The fitness value represents the slime mold. Indicates the first The location of each slime mold, Indicates the first The distance between the location of each slime mold and the predicted location of the incoming roll material is used to encourage slime molds to move closer to the coordinates of the incoming roll material. This indicates whether the location of the slime mold contains other rolls of material in that storage location. If so, the index of the roll of material in that location is recorded in the set. middle, Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the radius of the roll material entering the warehouse. Indicates the magnitude of suspension error. Indicates the first The slime mold represents the roll material entering the warehouse. The height difference between the lowest point of the coordinate system and the shelf. This indicates the shelving in this warehouse location. coordinate, Indicates the first The slime mold represents the roll material entering the warehouse. The lowest point of the coordinates and the first storage location One roll of material The height difference of the highest points of the coordinates Indicates the first A slime mold coordinate, This indicates the storage location One roll coordinate, Storage location The radius of each roll of material, Indicates the magnitude of the collision error. Indicates the first The boundary of the roll material represented by each slime mold in the warehouse is related to the first location in this warehouse. The edge bonding condition of each roll material.

[0144] By adopting the aforementioned design scheme, the beneficial effects of the present invention are as follows: This application uses a physical information neural network with added physical constraints adapted to the environment of the warehouse roll material to predict the roll radius of the warehouse roll material, and works in conjunction with the digital twin management system to realize the roll diameter prediction and coordinate calibration under physical rule constraints, thereby improving the consistency and usability of the positioning results in the three-dimensional warehousing scenario.

[0145] By obtaining the initial position coordinates of the incoming roll material through the positioning algorithm and fusing them with the predicted radius, a geometric entity and initial posture consistent with the real roll material can be quickly established in the digital twin platform. This provides accurate size boundaries and spatial references for subsequent physical rule legality judgments, such as suspension and collision, and provides a reasonable initial solution for position calibration / correction. This reduces the search range and number of iterations, lowers the risk of misjudgment and positioning drift, and improves the stability and accuracy of overall positioning calibration.

[0146] Compared to the traditional slime mold algorithm, the improved slime mold algorithm incorporates an initial position prior, which makes the algorithm more biased towards the initial position of the fabric in the later stages and gives the initial position greater weight. This is more in line with the environment of coordinate fine-tuning, and the improved slime mold algorithm is also more arbitrary in convergence than the original algorithm. Attached Figure Description

[0147] Figure 1 This diagram illustrates why the legality of the positioning of this application was not determined.

[0148] Figure 2 This is a schematic diagram of the roll material coordinate calibration for this application;

[0149] Figure 3 This is a schematic diagram comparing the aggregation of slime mold communities in the improved algorithm of this application. Detailed Implementation

[0150] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0151] The terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0152] A method for calibrating the position of a flexible roll material includes the following steps performed sequentially:

[0153] S1: Obtain the physical information neural network, obtain training samples, pre-train the physical information neural network, add physical constraints adapted to the environment of the warehouse roll material to the loss function, obtain the total loss function, use the total loss function to backpropagate and adjust the trained physical information neural network, obtain the prediction network, use the prediction network to predict the roll radius of the warehouse roll material, and obtain the prediction radius.

[0154] In this embodiment, the physical information neural network is preferably a PINN (MLP-PINN / FC-PINN) constructed using a multilayer perceptron or a fully connected feedforward network (MLP) as a function approximator; alternatively, residual network PINN (Res-PINN / ResNet-PINN), adaptive weight PINN (SA-PINN), or LSTM-PINN structures can also be used. All of these structures can introduce physical constraint terms into the supervised loss, outputting a predicted value for the roll radius of the incoming roll material.

[0155] In this embodiment, the training sample includes the type, length and weight of the roll material in the warehouse, and the physical constraints include weight ratio consistency, monotonicity prior and second-order curvature smoothness prior.

[0156] The calculation steps for the total loss function in step S1 are as follows:

[0157] S1-1: Consistency in weight ratios is based on the assumption that the specific density is unknown. Under the condition of the volume / mass conservation relationship of the roll material, the mass of each section of the roll material is... The corresponding cross-sectional increment should maintain a fixed proportion, and the loss function for the consistency of this weight ratio is:

[0158] ;

[0159] in, Indicates the quality of the roll material. Indicates the number of samples. This indicates the predicted radius, in meters. Indicates the radius of the roll of material entering the warehouse. Indicates the first indivual;

[0160] S1-2: The monotonicity prior is based on the length of the incoming roll material. After sorting, the corresponding prediction radius It should be monotonically non-decreasing; here, monotonically non-decreasing means that the latter term is greater than or equal to the former term.

[0161] Will One sample, for the length of the roll material entering the warehouse. After sorting in ascending order, we obtain the sorted index. :

[0162] ;

[0163] The loss function for monotonic priors is:

[0164] ;

[0165] in, The radius is the corresponding radius after sorting the lengths of the incoming rolls. Indicates the first indivual;

[0166] S1-3: Second-order curvature smoothing prior, based on the length of the roll material... In the ascending-order sorted sequence, the predicted radius is smoothed. The loss function for the second-order curvature smoothing prior is:

[0167] ;

[0168] S1-4: The total loss function is expressed by the following formula:

[0169] ;

[0170] ;

[0171] in, For the first The true value of the radius of each training sample. To monitor errors, , , These are the weighting coefficients for each loss function.

[0172] The predicted radius obtained in step S1 is used together with the initial position coordinates output in step S2 to determine the geometric entity size of the roll material in the digital twin platform; and serves as the size boundary parameter for the legality judgment (physical rules such as suspension / collision) in step S3 and the correction search (fitness and constraint calculation) in step S4, thereby ensuring that the calibration results meet the physical feasibility and improving the stability and accuracy of the positioning calibration.

[0173] Predicting the radius of the incoming roll material allows for the construction of a complete fabric model on the platform, based on the predicted radius and initial position. This model is then used to determine and calibrate the initial coordinates of the fabric, and to map the fabric model to the digital twin platform.

[0174] S2: Using a positioning algorithm, the initial position coordinates of the incoming roll material are obtained and combined with the predicted radius to map the incoming roll material to the digital twin platform;

[0175] In this embodiment, RFID readers are installed around the shelf, and RFID tags are affixed to the roll material. By reading the RFID signal on the RFID tag and combining it with the positioning algorithm, the shelf ID where the roll material is located and its two-dimensional coordinates relative to the shelf can be obtained. Before transmitting the coordinates to the digital twin platform, the two-dimensional coordinates need to be converted into three-dimensional coordinates. The conversion formula is as follows:

[0176] ;

[0177] in, This represents the predicted coordinates of the initial position of the incoming roll material obtained by the positioning algorithm. This indicates the coordinates of the incoming roll material after it has been mapped to the digital twin platform. The coordinate point is located at the center of the circle on the bottom surface of the cylindrical roll material. This indicates that the location is in the digital twin. The coordinates of the axis.

[0178] The initial position coordinates of the incoming roll material are obtained by the positioning algorithm and fused with the predicted radius obtained in step S1. This enables the rapid establishment of a geometric entity and initial posture consistent with the real roll material in the digital twin platform. This provides accurate dimensional boundaries and spatial references for the physical rule legality judgment in subsequent step S3, such as suspension and collision, and provides a reasonable initial solution for position calibration / correction in step S4. This reduces the search range and number of iterations, lowers the risk of misjudgment and positioning drift, and improves the stability and accuracy of overall positioning calibration.

[0179] S3: Perform a validity check on the initial location of the incoming roll material according to preset physical rules, such as... Figures 1-2 As shown, if the legality check fails, proceed to step S4; otherwise, proceed to step S5.

[0180] In step S3, the specific steps for verifying the legality of the initial location of the incoming roll material are as follows:

[0181] S3-1: Perform suspension detection on the incoming roll material based on Euclidean spatial measurement. The bounding box of the incoming roll material in the digital twin platform is:

[0182] ;

[0183] The center of the bottom support point of the incoming roll material is The coordinates of the two ends of the shelf in the warehouse where the incoming roll material is located are respectively , ,

[0184] in, This represents the horizontal coordinate of the shelving unit in this storage location. and This indicates that the incoming roll material is in the digital twin platform. Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis;

[0185] S3-2: Determine if the incoming roll material is in a suspended state: Project the bottom center point of the incoming roll material onto the shelf. Plane, projection point is ;

[0186] Define the shelf vector and the projection point vector, and calculate the projection ratio of the projection point vector onto the shelf vector using the following formula:

[0187] ;

[0188] ;

[0189] like The projection point will fall inside the shelf. This indicates that the shelf support height threshold is an empirical value; otherwise, the ratio of the vertical height of the bottom center of the incoming roll material to the shelf height is calculated using the following formula:

[0190] ;

[0191] like If the incoming roll material is supported by the shelf, it will not be suspended in mid-air. This indicates that the shelf support height threshold is an empirical value and can be set according to actual needs. It is generally set to 0.01m. Otherwise, proceed to step S3-3.

[0192] S3-3: The following formula is used to determine whether the incoming roll material is supported by other roll materials:

[0193] ;

[0194] ;

[0195] ;

[0196] in, This indicates the location of the storage unit. The size of the wrapping box for each roll of material. This indicates the location of the storage unit. The center point of the roll material is at Projected coordinates of the plane The radius of the roll of material to be stored. , and They represent, , and They represent respectively;

[0197] Calculate the horizontal distance between the bottom center of the incoming roll material and other roll materials:

[0198] ;

[0199] in, and Is the incoming roll material in The coordinates of the projection point on the plane, This indicates that the incoming roll material is related to the first item in the warehouse location. Each roll of material Distance between points projected onto the plane;

[0200] If satisfied and , This is the tolerance value for horizontal difference. The threshold height for the support of the incoming roll material. This is an empirical value, generally set to 0.01m; then it is considered that the incoming roll material is stored in the warehouse at the location of the first... Supported by the roll material, without any suspension, proceed to step S3-4 to perform multi-level collision detection;

[0201] S3-4: Employs an axis-aligned bounding box algorithm for rapid detection, searching through all rolls in this storage location to identify the indices of rolls that may collide with the incoming roll:

[0202] ;

[0203] ;

[0204] in, and These respectively represent the incoming roll material and the first one in the storage location. A box surrounding a roll of material. For roll materials that need to be further checked for collisions after rapid initial screening by the axis alignment box;

[0205] like If the length is not empty, the following formula is used to determine... Are there any rolls in the collection that collided with the incoming roll?

[0206] ;

[0207] ;

[0208] in, Indicates the center of the roll material is at The projection on the plane, then and These respectively indicate the condition of the incoming roll material and the roll material after initial screening. Projection on a plane Indicates the radius of the roll material, then and These represent the radii of the incoming roll material and the roll material after initial screening, respectively.

[0209] For each item It is necessary to determine whether the following collision determination conditions are met:

[0210] ;

[0211] in, This is the tolerance value for the radius;

[0212] If the collision determination conditions are not met, it is considered that a collision has occurred. That is, when the center distance between two rolls (or rolls and obstacles) in the plane projection is less than or equal to the sum of their equivalent radii (after deducting / considering tolerances), it is determined that a collision has occurred or there is a risk of collision; otherwise, it is considered that no collision has occurred and the safe distance is met.

[0213] S4: The Slime Mould Algorithm with Original-Position Prior (SMA-OP) is used to search for target positions that conform to physical rules near the initial position. In the SMA-OP algorithm with initial position prior, the algorithm that approaches the origin in the local search mode of the original SMA algorithm is replaced by a mode that scales the current coordinates and adds the initial coordinate weight of the roll material. This avoids all individuals being pulled towards the origin and instead gathers them in the region where the true global optimum is located. An improved optimization algorithm that conforms to the search for the roll material coordinates is constructed. The specific steps are as follows:

[0214] Each slime mold location represents a suitable position for placing the incoming roll material. Based on the traditional Slime Mould Algorithm with Original (SMA), the local search is modified by incorporating an initial position prior. The specific formula is as follows:

[0215] ;

[0216] ;

[0217] ;

[0218] in, The maximum number of iterations, Indicates the first iteration Indicates taking 0 to... Random numbers between For the first The location of the slime mold, The initial two-dimensional coordinates representing the location of the incoming roll material. This represents the mixed coefficient vector of the slime mold algorithm in the local search mode. This represents the convergence factor that varies linearly with iteration. , , , This represents the initial three-dimensional coordinates of the incoming roll material obtained by the positioning algorithm;

[0219] The fitness value of the population is calculated iteratively using the fitness function of this slime mold algorithm.

[0220] In this embodiment, the slime mold algorithm with initial position prior has enhanced ability to approximate the original coordinate point in the later stage, which is more in line with the application scenario of micro-calibration of roll material storage coordinates. The fitness function of this slime mold algorithm is:

[0221] ;

[0222] ;

[0223] ;

[0224] ;

[0225] in, The fitness value represents the slime mold. Indicates the first The location of each slime mold, Indicates the first The distance between the location of each slime mold and the predicted location of the incoming roll material is used to encourage slime molds to move closer to the coordinates of the incoming roll material. This indicates whether the location of the slime mold contains other rolls of material in that storage location. If so, the index of the roll of material in that location is recorded in the set. middle, Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the radius of the roll material entering the warehouse. Indicates the magnitude of suspension error. Indicates the first The slime mold represents the roll material entering the warehouse. The height difference between the lowest point of the coordinate system and the shelf. This indicates the shelving in this warehouse location. coordinate, Indicates the first The slime mold represents the roll material entering the warehouse. The lowest point of the coordinates and the first storage location One roll of material The height difference of the highest points of the coordinates Indicates the first A slime mold coordinate, This indicates the storage location One roll coordinate, Storage location The radius of each roll of material, Indicates the magnitude of the collision error. Indicates the first The boundary of the roll material represented by each slime mold in the warehouse is related to the first location in this warehouse. The edge bonding condition of each roll of material;

[0226] like Figure 3 As shown, the fitness function and population are repeatedly calculated until the preset number of iterations is reached, and the optimal result of the slime mold algorithm is output, which is the target location of the roll material to be put into storage.

[0227] In the local optimization phase of SMA, the algorithm gradually shrinks the individual update step size to refine the current optimal solution. However, this shrinkage behavior is based on an "oscillation factor" that monotonically decreases with the number of iterations. As it gradually approaches its maximum value, the subsequent update weights also tend towards the origin. This means that this "oscillation factor" tends to 0. Although this design can refine the search to some extent, in the context of fabric storage, it can lead to invalid searches, making it difficult for the algorithm to escape local optima and ultimately find the true global optimum. The improved slime mold algorithm incorporates an initial position prior, making the algorithm more biased towards the initial fabric position in the later stages and giving the initial position greater weight, which is more suitable for the environment of coordinate fine-tuning. The improved slime mold algorithm also converges more arbitrarily than the original algorithm.

[0228] S5: Remap the incoming roll material to the digital twin platform based on the coordinates of the target location.

[0229] This embodiment also provides a warehouse management system for flexible roll materials to implement the above method.

[0230] A warehouse management system for flexible roll materials includes a roll material radius prediction module, a roll material positioning module, a roll material positioning validity judgment module, a roll material positioning correction module, and a roll material position display module;

[0231] The roll radius prediction module is used to acquire physical information neural network and training samples, and to pre-train the physical information neural network using the training samples. Physical constraints adapted to the environment of the roll material entering the warehouse are added to the loss function of the physical information neural network to obtain the total loss function. The total loss function is used to backpropagate and adjust the trained physical information neural network to obtain the prediction network. The prediction network is used to predict the roll radius of the roll material entering the warehouse to obtain the predicted radius.

[0232] In this embodiment, the training sample includes the type, length and weight of the roll material in the warehouse, and the physical constraints include weight ratio consistency, monotonicity prior and second-order curvature smoothness prior.

[0233] The steps for calculating the total loss function are as follows:

[0234] Consistency in weight ratios is based on the assumption that the specific density is unknown. Under the condition of the volume / mass conservation relationship of the roll material, the mass of each section of the roll material is... The corresponding cross-sectional increment should maintain a fixed proportion, and the loss function for the consistency of this weight ratio is:

[0235] ;

[0236] in, Indicates the quality of the roll material. Indicates the number of samples. This indicates the predicted radius, in meters. Indicates the radius of the roll of material entering the warehouse. Indicates the first indivual,;

[0237] The monotonicity prior is the length of the incoming roll material. After sorting, the corresponding prediction radius It should be monotonically non-decreasing; here, monotonically non-decreasing means that the latter term is greater than or equal to the former term.

[0238] Will One sample, for the length of the roll material entering the warehouse. After sorting in ascending order, we obtain the sorted index. :

[0239] ;

[0240] The loss function for monotonic priors is:

[0241] ;

[0242] in, The radius is the corresponding radius after sorting the lengths of the incoming rolls. Indicates the first indivual,;

[0243] The second-order curvature smoothing prior is based on the length of the roll material. In the ascending-order sorted sequence, the predicted radius is smoothed. The loss function for the second-order curvature smoothing prior is:

[0244] ;

[0245] in, The radius corresponding to the length of the incoming roll material after sorting.

[0246] The total loss function is expressed by the following formula:

[0247] ;

[0248] ;

[0249] in, For the first The true value of the radius of each training sample. To monitor errors, , , These are the weighting coefficients for each loss function.

[0250] The roll material positioning module is used to obtain the initial position coordinates of the inbound roll material using a positioning algorithm and combine the predicted radius to map the inbound roll material onto the roll material position display module;

[0251] In this embodiment, RFID readers are installed around the shelf, and RFID tags are affixed to the roll material. By reading the RFID signal on the RFID tag and combining it with the positioning algorithm, the shelf ID where the roll material is located and its two-dimensional coordinates relative to the shelf can be obtained. Before transmitting the coordinates to the digital twin platform, the two-dimensional coordinates need to be converted into three-dimensional coordinates. The conversion formula is as follows:

[0252] ;

[0253] in, This represents the predicted coordinates of the initial position of the incoming roll material obtained by the positioning algorithm. This indicates the coordinates of the incoming roll material after it has been mapped to the digital twin platform. The coordinate point is located at the center of the circle on the bottom surface of the cylindrical roll material. This indicates that the location is in the digital twin. The coordinates of the axis.

[0254] The roll material positioning legality judgment module is used to judge the legality of the initial position of the incoming roll material according to preset physical rules;

[0255] The specific steps for verifying the legality of the initial location of the incoming roll material are as follows:

[0256] The floating detection of the incoming roll material was performed based on Euclidean spatial measurement. The bounding box of the incoming roll material in the digital twin platform is as follows:

[0257] ;

[0258] The center of the bottom support point of the incoming roll material is The coordinates of the two ends of the shelf in the warehouse where the incoming roll material is located are respectively , ,

[0259] in, This represents the horizontal coordinate of the shelving unit in this storage location. and This indicates that the incoming roll material is in the digital twin platform. Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis;

[0260] Determine if the incoming roll material is in a suspended state: Project the center point of the bottom of the incoming roll material onto the shelf. Plane, projection point is ;

[0261] Define the shelf vector and the projection point vector, and calculate the projection ratio of the projection point vector onto the shelf vector using the following formula:

[0262] ;

[0263] ;

[0264] like The projection point will fall inside the shelf. This indicates that the shelf support height threshold is an empirical value; otherwise, the ratio of the vertical height of the bottom center of the incoming roll material to the shelf height is calculated using the following formula:

[0265] ;

[0266] like If the incoming roll material is supported by the shelf, it will not be suspended in mid-air. This indicates that the shelf support height threshold is an empirical value and can be set according to actual needs. It is generally set to 0.01m. Otherwise, the following formula is used to determine whether the incoming roll material is supported by other roll materials:

[0267] ;

[0268] ;

[0269] ;

[0270] in, This indicates the location of the storage unit. The size of the wrapping box for each roll of material. This indicates the location of the storage unit. The center point of the roll material is at Projected coordinates of the plane The radius of the roll of material to be stored. , and They represent, , and They represent respectively;

[0271] Calculate the horizontal distance between the bottom center of the incoming roll material and other roll materials:

[0272] ;

[0273] in, and Is the incoming roll material in The coordinates of the projection point on the plane, This indicates that the incoming roll material is related to the first item in the warehouse location. Each roll of material Distance between points projected onto the plane;

[0274] If satisfied and , This is the tolerance value for horizontal difference. The threshold for the support height of the incoming roll material is an empirical value, generally set at 0.01m; therefore, it is considered that the incoming roll material is supported by the first unit in the storage location. Supported by the roll material, without any suspension, proceed to step S3-4 to perform multi-level collision detection;

[0275] A bounding box algorithm with axis alignment is used for rapid detection to select the index of the roll that may collide with the incoming roll from all rolls in this storage location:

[0276] ;

[0277] ;

[0278] in, and These respectively represent the incoming roll material and the first one in the storage location. A box surrounding a roll of material. For roll materials that need to be further checked for collisions after rapid initial screening by the axis alignment box;

[0279] like If the length is not empty, the following formula is used to determine... Are there any rolls in the collection that collided with the incoming roll?

[0280] ;

[0281] ;

[0282] in, Indicates the center of the roll material is at The projection on the plane, then and These respectively indicate the condition of the incoming roll material and the roll material after initial screening. Projection on a plane Indicates the radius of the roll material, then and These represent the radii of the incoming roll material and the roll material after initial screening, respectively.

[0283] For each item It is necessary to determine whether the following collision determination conditions are met:

[0284] ;

[0285] in, This is the tolerance value for the radius.

[0286] The roll material positioning correction module uses a slime mold algorithm with an initial position prior to search for a target position that conforms to physical rules near the initial position. The specific operation process is as follows:

[0287] Each slime mold location represents a suitable location for placing the incoming roll material. Based on the traditional slime mold algorithm, the local search is modified by incorporating an initial location prior, as shown in the following formula:

[0288] ;

[0289] ;

[0290] ;

[0291] in, The maximum number of iterations, Indicates the first iteration Indicates taking 0 to... Random numbers between For the first The location of the slime mold, The initial two-dimensional coordinates representing the location of the incoming roll material. This represents the mixed coefficient vector of the slime mold algorithm in the local search mode. This represents the convergence factor that varies linearly with iteration. , , , This represents the initial three-dimensional coordinates of the incoming roll material obtained by the positioning algorithm;

[0292] The fitness value of the population is calculated iteratively using the fitness function of this slime mold algorithm.

[0293] In this embodiment, the slime mold algorithm with initial position prior has enhanced ability to approximate the original coordinate point in the later stage, which is more in line with the application scenario of micro-calibration of roll material storage coordinates. The fitness function of this slime mold algorithm is:

[0294] ;

[0295] ;

[0296] ;

[0297] ;

[0298] in, The fitness value represents the slime mold. Indicates the first The location of each slime mold, Indicates the first The distance between the location of each slime mold and the predicted location of the incoming roll material is used to encourage slime molds to move closer to the coordinates of the incoming roll material. This indicates whether the location of the slime mold contains other rolls of material in that storage location. If so, the index of the roll of material in that location is recorded in the set. middle, Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the radius of the roll material entering the warehouse. Indicates the magnitude of suspension error. Indicates the first The slime mold represents the roll material entering the warehouse. The height difference between the lowest point of the coordinate system and the shelf. This indicates the shelving in this warehouse location. coordinate, Indicates the first The slime mold represents the roll material entering the warehouse. The lowest point of the coordinates and the first storage location One roll of material The height difference of the highest points of the coordinates Indicates the first A slime mold coordinate, This indicates the storage location One roll coordinate, Storage location The radius of each roll of material, Indicates the magnitude of the collision error. Indicates the first The boundary of the roll material represented by each slime mold in the warehouse is related to the first location in this warehouse. The edge bonding condition of each roll of material;

[0299] Repeat the calculation of the fitness function and update the population until the preset number of iterations is reached, and output the optimal result of the slime mold algorithm, which is the target location of the rolled material to be put into storage.

[0300] The roll material location display module is used to display the target location of the incoming roll material on the system front end.

[0301] In summary, this application employs a physical information neural network with added physical constraints adapted to the inbound roll material environment to predict the roll material radius, and collaborates with a digital twin management system to achieve roll diameter prediction and coordinate calibration under physical rule constraints, thereby improving the consistency and usability of positioning results in a three-dimensional warehousing scenario.

[0302] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for calibrating the position of a flexible roll material, characterized in that: The steps are as follows, performed sequentially: S1: Obtain the physical information neural network, obtain training samples, pre-train the physical information neural network, add physical constraints adapted to the environment of the warehouse roll material to the loss function, obtain the total loss function, use the total loss function to backpropagate and adjust the trained physical information neural network, obtain the prediction network, use the prediction network to predict the roll radius of the warehouse roll material, and obtain the prediction radius. S2: Using a positioning algorithm, the initial position coordinates of the incoming roll material are obtained and combined with the predicted radius to map the incoming roll material to the digital twin platform; S3: Perform a validity check on the initial position of the incoming roll material according to the preset physical rules. If the validity check fails, proceed to step S4; otherwise, proceed to step S5. S4: The slime mold algorithm with initial position prior is used to search for target positions that conform to physical rules near the initial position. The specific steps are as follows: Each slime mold location represents a suitable location for placing the incoming roll material. Based on the traditional slime mold algorithm, the local search is modified by incorporating an initial location prior, as shown in the following formula: ； ； ； in, The maximum number of iterations, Indicates the first iteration Indicates taking 0 to... Random numbers between For the first The location of the slime mold, The initial two-dimensional coordinates representing the location of the incoming roll material. This represents the mixed coefficient vector of the slime mold algorithm in the local search mode. This represents the convergence factor that varies linearly with iteration. , , , This represents the initial three-dimensional coordinates of the incoming roll material obtained by the positioning algorithm; The fitness value of the population is calculated iteratively using the fitness function of this slime mold algorithm. Repeat the calculation of the fitness function and update the population until the preset number of iterations is reached, and output the optimal result of the slime mold algorithm, which is the target location of the rolled material to be put into storage. S5: Remap the incoming roll material to the digital twin platform based on the coordinates of the target location.

2. The position positioning and calibration method for flexible roll material as described in claim 1, characterized in that: In step S1, the training sample includes the type, length and weight of the roll material in the warehouse, and the physical constraints include weight ratio consistency, monotonicity prior and second-order curvature smoothness prior. The calculation steps for the total loss function in step S1 are as follows: S1-1: The loss function for the consistency of this weight ratio is: ； in, Indicates the quality of the roll material. Indicates the number of samples. Indicates the predicted radius. Indicates the radius of the roll material in the warehouse; S1-2: The monotonicity prior is based on the length of the incoming roll material. After sorting, the corresponding prediction radius It should be monotonous and undiminished; One sample, for the length of the roll material entering the warehouse. After sorting in ascending order, we obtain the sorted index. : ； The loss function for monotonic priors is: ； in, The radius corresponding to the length of the incoming roll material after sorting. S1-3: The loss function for the second-order curvature smoothing prior is: ； S1-4: The total loss function is expressed by the following formula: ； ； in, For the first The true value of the radius of each training sample. To monitor errors, , , These are the weighting coefficients for each loss function.

3. The position positioning and calibration method for flexible roll materials as described in claim 2, characterized in that: In step S3, the specific steps for verifying the legality of the initial location of the incoming roll material are as follows: S3-1: Perform suspension detection on the incoming roll material based on Euclidean spatial measurement. The bounding box of the incoming roll material in the digital twin platform is: ； The center of the bottom support point of the incoming roll material is The coordinates of the two ends of the shelf in the warehouse where the incoming roll material is located are respectively , , in, This represents the horizontal coordinate of the shelving unit in this storage location. and This indicates that the incoming roll material is in the digital twin platform. Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis, and Indicates in Minimum and maximum values ​​on the axis; S3-2: Project the bottom center point of the incoming roll material onto the shelf. Plane, projection point is ； Define the shelf vector and the projection point vector, and calculate the projection ratio of the projection point vector onto the shelf vector using the following formula: ； ； ； like The projection point will fall inside the shelf. Indicates the threshold height for shelf support. This is an empirical value; otherwise, use the following formula to calculate the ratio of the vertical height of the bottom center of the incoming roll material to the shelf height: ； like If the incoming roll material is supported by the shelf, then proceed to step S3-3; S3-3: The following formula is used to determine whether the incoming roll material is supported by other roll materials: ； ； ； in, This indicates the first [unit / location] of the storage location. The size of the wrapping box for each roll of material. This indicates the first [unit / location] of the storage location. The center point of the roll material is at Projected coordinates of the plane The radius of the roll of material to be stored; Calculate the horizontal distance between the bottom center of the incoming roll material and other roll materials: ； in, and Is the incoming roll material in The coordinates of the projection point on the plane, This indicates that the incoming roll material is related to the first item in the warehouse location. Each roll of material Distance between points projected onto the plane; If satisfied and , This is the tolerance value for horizontal difference. The threshold height for the support of the incoming roll material. If the value is empirical, then it is considered that the incoming roll material is stored in the warehouse location. Supported by the roll material, without any suspension, proceed to step S3-4 to perform multi-level collision detection; S3-4: Using an axis-aligned bounding box algorithm for rapid detection, the index of the roll that collides with the incoming roll is selected from all rolls in this storage location: ； ； in, and These respectively represent the incoming roll material and the first one in the storage location. A box surrounding a roll of material. For roll materials that need to be further checked for collisions after rapid initial screening by the axis alignment box; like If the length is not empty, the following formula is used to determine... Are there any rolls in the collection that collided with the incoming roll? ； ； in, Indicates the center of the roll material is at The projection on the plane, then and These respectively indicate the condition of the incoming roll material and the roll material after initial screening. Projection on a plane Indicates the radius of the roll material, then and These represent the radii of the incoming roll material and the roll material after initial screening, respectively. For each item It is necessary to determine whether the following collision determination conditions are met: ； in, This is the tolerance value for the radius.

4. The position positioning and calibration method for flexible roll material as described in claim 3, characterized in that: In step S4, the fitness function of the slime mold algorithm is: ； ； ； ； in, The fitness value represents the slime mold. Indicates the first The location of each slime mold, Indicates the first The distance between the location of each slime mold and the predicted location of the incoming roll material is used to encourage slime molds to move closer to the coordinates of the incoming roll material. This indicates whether the location of the slime mold is surrounded by other rolls of material in that storage location. If so, the index of the roll of material in that location is recorded in the set. middle, Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the radius of the roll material entering the warehouse. Indicates the magnitude of suspension error. Indicates the first The slime mold represents the roll material entering the warehouse. The height difference between the lowest point of the coordinate system and the shelf. This indicates the shelving in this warehouse location. coordinate, Indicates the first The slime mold represents the roll material entering the warehouse. The lowest point of the coordinates and the first storage location One roll of material The height difference of the highest points of the coordinates Indicates the first A slime mold coordinate, This indicates the location of the storage unit. One roll of material coordinate, Indicates the location of the storage unit. The radius of each roll of material, Indicates the magnitude of the collision error. Indicates the first The boundary of the roll material represented by each slime mold in the warehouse is related to the first location in this warehouse. The edge bonding condition of each roll material.

Images